Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I want to show that

$ Y(t)=\cos(At)$ and $Y(t)=\sin(At)$ satisfy the equation

$$Y''+A^2Y=0 $$

subjected to the initial conditions $Y(0)=I, Y'(0)=0$ and $Y(0)=0, Y'(0)=I$ respictively

where $A$ is an constant $n \times n$ matrix..

share|cite|improve this question
Something is strange here...if $Y(t)=\cos(At)$, then $Y(0)$ is never going to be the identity matrix. It will be a matrix full of ones. That is, assuming $t$ is a scalar...? – icurays1 Jan 14 '13 at 23:48
@icurays1 It's not element-wise cosine. See the Taylor expansion below in Mhenni's answer. – John Moeller Jan 14 '13 at 23:52
Well. That's embarrassing. – icurays1 Jan 14 '13 at 23:53
up vote 1 down vote accepted

If $$X(t):=\cos(A t):=\sum_{k=0}^\infty{\cos(k\pi/2)\over k!} A^k t^k$$ then $X(0)=I$ and $$X'(t)=\sum_{k=1}^\infty{\cos(k\pi/2)\over k!} k A^k t^{k-1}=-A\ \sum_{k=0}^\infty{\sin(k\pi/2)\over k!} A^k t^k=-A\ \sin(At)\ ;$$ in particular $X'(0)=0$.

Furthermore, if $$Y(t):=\sin(A t):=\sum_{k=1}^\infty{\sin(k\pi/2)\over k!} A^k t^k$$ then $Y(0)=0$ and $$Y'(t)=\sum_{k=1}^\infty{\sin(k\pi/2)\over k!} k A^k t^{k-1}=A\ \sum_{k=0}^\infty{\cos(k\pi/2)\over k!} A^k t^k=A\ \cos(At)\ ;$$ in particular $Y'(0)=A$.

It follows that $X''(t)=-A Y'(t)=-A^2 X(t)$; so $X(\cdot)$ satisfies the given differential equation as well as the initial conditions $X(0)=I$, $X'(0)=0$.

For $Y(\cdot)$ it's different: This function satisfies the differential equation as well as the initial condition $Y(0)=0$. But if we want $Y'(0)=I$ instead of $Y'(0)=A$ we have to replace $Y(\cdot)$ by the new solution $$\hat Y(t):=A^{-1}Y(t)=A^{-1}\sin(A t)\ .$$ When $A$ happens to be singular we seem to be in trouble. But there is remedy: Since the sine series begins with the linear term we can write $\hat Y(t)$ without explicit appearance of $A^{-1}$: $$\hat Y(t)=\sum_{k=1}^\infty{\sin(k\pi/2)\over k!} A^{k-1} t^k= t\ {\rm sinc}(A t)\ .$$

share|cite|improve this answer


1) Use the power series representation

$$ Y(t)=\sin(At) = \sum_{k=0}^{\infty}(-1)^kA^{2k+1}\frac{t^{2k+1}}{(2k+1)!}, $$

and same for $\cos(At)$.

2) If $Y_1$ and $Y_2$ are solutions of the ode, then

$$Y := c_1Y_1+c_2 Y_2$$

is a solution too.


$$ Y(t)=\sin(At) = \sum_{k=0}^{\infty}(-1)^kA^{2k+1}\frac{t^{2k+1}}{(2k+1)!} $$

$$\implies Y'(t)=\sum_{k=0}^{\infty}(-1)^kA^{2k+1}(2k+1)\frac{t^{2k}}{(2k+1)!} $$

$$\implies Y''(t)=\sum_{k=1}^{\infty}(-1)^kA^{2k+1}(2k+1)(2k)\frac{t^{2k-1}}{(2k+1)!}=\sum_{k=1}^{\infty}(-1)^kA^{2k+1}\frac{t^{2k-1}}{(2k-1)!}. $$

Shifting the index in the above sum gives

$$ \sum_{k=0}^{\infty}(-1)^{k+1}A^{2(k+1)+1}\frac{t^{2(k+1)-1}}{(2(k+1)-1)!}= -A^2 \sum_{k=0}^{\infty}(-1)^{k}A^{2k+1}\frac{t^{2k+1}}{(2k+1)!}$$

$$ \implies Y''(t) = -A^2 \sin(At). $$

share|cite|improve this answer
How can you say $$Y'(t)=\sum_{k=0}^{\infty}(-1)^kA^{2k+1}(2k+1)\frac{t^{2k}}{(2k+1)!} \implies Y''(t)=\sum_{k=1}^{\infty}(-1)^kA^{2k+1}(2k+1)(2k)\frac{t^{2k-1}}{(2k+1)!} $$ How does $k=0$ change into $k=1$ ? – Applied mathematician Jan 15 '13 at 17:38
Hmm... I see, if k=0, then the whole term vanishes. – Applied mathematician Jan 15 '13 at 18:22
What do I have to do with the initial conditions? I can show that $Y''(t)=0A^2sin(At)$... – Applied mathematician Jan 15 '13 at 20:57
You need to find the constants $c_1$ and $c_2$. – Mhenni Benghorbal Jan 15 '13 at 22:46
How can you do that? – Applied mathematician Jan 16 '13 at 12:57

I found the shifting of indices to be highly confusing so I propose a method that does not rely on that.

We will first show that $Y(t)=\cos (At)$ solves the initial value problem $Y''+A^2 Y=0$ where $Y(0)=I$ and $Y'(0)=0$. Because $\cos At$ is defined as the infinite sum $\sum\limits_{k=0}^{\infty} (-1)^{k}\frac{A^{2k}t^{2k}}{(2k)!}$ and differentiation is linear we can differentiate term by term to get:

\begin{eqnarray*} Y''(t)&=&\frac{d^2}{dt^2}\left[ I -\frac{A^2t^2}{2!}+\frac{A^4t^4}{4!}-\frac{A^6t^6}{6!}+\ldots \right]\\ &=&\left[ -2\frac{A^2}{2!} +4\cdot 3 \frac{A^4 t^2}{4!} -6\cdot 5 \frac{A^6 t^4}{6!} +\ldots \right]\\ &=&-A^2\left[ I -\frac{A^2t^2}{2!}+\frac{A^4t^4}{4!}-\frac{A^6t^6}{6!}+\ldots \right]\\ &=& -A^2 \sum\limits_{k=0}^{\infty} (-1)^k \frac{A^{2k}t^{2k} }{(2k)!}=-A^2Y(t) \end{eqnarray*}

So $Y(t)$ satisfies $ Y''+A^2Y=0$. Also $Y(0)=\cos(0)=\sum\limits_{k=0}^{\infty} (-1)^{k}\frac{0^{2k}}{(2k)!}=I$ (The $0^{2k}$ in the summation is the zero matrix $[A\cdot t]_{t=0}=0$ and $0^0=I$, all higher powers of $0$ are of course $0$). And for $Y'(0)$ we find

\begin{eqnarray*} Y'(0)&=&\frac{d}{dt}\left[ I -\frac{A^2t^2}{2!}+\frac{A^4t^4}{4!}-\frac{A^6t^6}{6!}+\ldots \right] _{t=0}\\ &=& \left[ -2\frac{A^20}{2!}+4\frac{A^40^3}{4!}-6\frac{A^60^5}{6!}+\ldots \right] =0 \end{eqnarray*} (here the $0^k$ is just the scalar $t=0$ )

For $Y(t)=\sin(At)$ with $\sin(At):=\sum\limits_{k=0}^{\infty} (-1)^{k}\frac{A^{2k+1}t^{2k+1}}{(2k+1)!}$ we have:

\begin{eqnarray*} Y''(t)&=&\frac{d^2}{dt^2}\left[ At -\frac{A^3t^3}{3!}+\frac{A^5t^5}{5!}-\frac{A^7t^7}{7!}+\ldots \right]\\ &=&\left[ -3\cdot 2\frac{A^3t}{3!} +5\cdot 4 \frac{A^5 t^3}{5!} -5\cdot 6 \frac{A^7 t^5}{7!} +\ldots \right]\\ &=&-A^2\left[ At -\frac{A^3t^3}{3!}+\frac{A^5t^5}{5!}-\frac{A^7t^7}{7!}+\ldots \right]\\ &=& -A^2 \sum\limits_{k=0}^{\infty} (-1)^{k}\frac{A^{2k+1}t^{2k+1}}{(2k+1)!}=-A^2Y(t) \end{eqnarray*}

So $Y(t)=\sin(At)$ also satisfies $Y''(t)+A^2Y(t)=0$. Here however $Y(0)=\sin( 0)=\sum\limits_{k=0}^{\infty} (-1)^{k}\frac{0^{2k+1}}{(2k+1)!}=0$ (again using $0$ as the zero matrix in the summation, and now the lowest power of $0$ is $0^1$ so here we have an infinite sum of zeros). And

\begin{eqnarray} Y'(0)&=&\frac{d}{dt}\left[ At -\frac{A^3t^3}{3!}+\frac{A^5t^5}{5!}-\frac{A^7t^7}{7!}+\ldots \right] _{t=0}\\ &=& \left[ A -3\frac{A^30^2}{3!}+5\frac{A^50^4}{5!}-7\frac{A^70^6}{7!}+\ldots \right] =A \end{eqnarray} I am guessing this is what you meant because even if we consider a scalar $A$ then $\left[\frac{d}{dt}\sin(At)\right]_{t=0}=A\cos(0)=A$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.